摘要
目标跟踪是机器视觉领域的一项重要技术。传统的CamShift目标跟踪算法具有时间复杂度低、运算速度快的优点,在简单背景下具有良好的跟踪效果。但当跟踪目标处于部分遮挡的复杂情况下时,容易出现目标丢失的情况,从而影响后续的跟踪。Kalman滤波算法在目标跟踪任务中,能够有效地预测跟踪目标下一时刻可能出现的位置,且算法简单方便。现将CamShift算法与Kalman滤波算法相结合来优化传统的CamShift算法,实验结果表明:优化后的算法不但保留了传统CamShift算法的优点,而且在一定程度上可以预测跟踪目标的行动轨迹,更好地实现目标跟踪任务,解决了传统CamShift算法在目标被部分遮挡的情况下容易出现的目标丢失的情况;同时,运算结果准确,基本没有预测误差。
The tracking of targets plays a crucial role in machine vision.The traditional CamShift algorithm for target tracking offers the benefits of low time complexity and speedy operation.,and has good tracking effect in simple background.However,when the tracking target is in the complex condition of partial occlusion,it is easy to lose the target,which will affect the follow-up tracking.The Kalman filter algorithm is capable of accurately forecasting the potential next position of the tracked target in target tracking tasks,and the algorithm is simple and convenient.In this paper,the traditional CamShift target tracking algorithm is optimized by combining CamShift algorithm and Kalman filter algorithm.The integrated algorithm not only preserves the strengths of the traditional CamShift algorithm during the tracking of moving objects,but also provides a certain level of prediction for the tracking target’s trajectory,thereby optimizing the algorithm.Experimental results demonstrate that the integrated algorithm effectively handles target tracking tasks,achieving precise tracking even when the target is partially obstructed.It solves the problem that the traditional CamShift algorithm is prone to target loss when the target is incomplete blocked.Meanwhile,its calculation results are accurate,the tracking effect is good,and the prediction is basically without error.
作者
李俊松
刘光宇
王帅
程远
周豹
赵恩铭
杨春丽
LI Junsong;YANG Chunli;WANG Shuai;CHENG Yuan;ZHOU Bao;ZHAO Enming;LIU Guangyu(School of Engineering,Dali University,Dali,Yunnan 671003,China;Ministry of Education Shanghai Jiao Tong University,Laboratory for Ocean Intelligent Equipment and Systems,Shanghai 200030,China;Information Department of the People’s Liberation Army Unit 32268,Dali,Yunnan 671003,China)
出处
《山东商业职业技术学院学报》
2024年第3期116-120,共5页
Journal of Shandong Institute of Commerce and Technology
基金
国家自然科学基金资助项目(NO.62065001)
2021年云南省科技厅科技计划项目(202101BA070001-054)
云南省中青年学术和技术带头人后备人才项目(202205AC160001)
海洋智能装备与系统教育部重点实验室开放基金项目(MIES-2023-02)。
